Design limits and solutions for very large wind turbines

More documents

Recommendations

Info

3 UpWind: Technology integration Generators The main objective was to fi nd the most suitable generator system for the large wind turbines expected by 2020. Currently, there are three main generator systems used in wind turbines: constant speed with a gearbox and a squirrel cage induction generator, variable speed with a gearbox and doubly-fed induction generators, and variable speed direct-drive generators without a gearbox. Each concept has its disadvantages. A first task was a thorough comparison of these and other possible generator concepts based on cost models and energy yield models. An overview of wind turbine technologies and a comparison of different generator topologies based on literature and market supply review was carried out. Promising direct-drive permanent magnet (PM) machines, which include the axial fl ux (AF), radial flux (RF) and transverse flux (TF) machines, have been surveyed in literature and compared based on the technical data and market aspects. In case of RFPM machines, it can be concluded that the machines are almost optimised electromagnetically, so that it is hard to reduce the active material weight and cost of the machines signifi cantly. The disadvantages of the AFPM machine must be solved, because it causes the machine cost to increase and manufacturing to be diffi cult. TFPM machines have disadvantages such as complicated construction and manufacturing, and low power factor, although the machines have advantages such as high force density and simple winding with low copper losses. The second task was to evaluate the most cost-effective wind turbine generator systems by applying design optimisation and numerical comparison. An integrated electromagnetic-structural optimisation strategy was developed to provide the best structural design for direct drive RFPM generators. The preliminary results of a structural optimisation highlighted the danger of not optimising the active and inactive material together – that is, the fact that a generator design that minimises active mass leads to a design that maximises inactive or structural mass. Traditionally in the design of the active mass, the air gap is kept as small as possible to optimise the electromagnetic performance. However, the integrated electromagnetic–structural optimisation strategy indicated that machines with a larger air gap will result in lower mass. The aspect ratio of the generator (ratio of length to air gap radius), depends upon the optimisation criteria. For minimum mass, large aspect ratios with a larger air gap is desirable, leading to a sausage shaped machine. If cost is key, then small aspect ratios or pancake machines are more desirable, because active mass decreases with radius, and this is the most expensive part of the generator. The optimisation strategies were also been applied to TFPM machine topologies. A third task was to identify the most suitable generator type for large direct-drive wind turbines based on electromagnetic analysis models for various confi gurations of PM machines. The surface mounted RFPM machine and four different fl ux-concentrating TFPM machines with single-winding were chosen for the comparative design, where a single-sided single-winding flux-concentrating TFPM generator was found to be most promising. The analysis models were verifi ed by experiments and finite element analyses of a down-scaled, linearised generator. In order to further reduce the active mass of TFPM machines, a confi guration with multiple-slots is proposed in order to shorten the length of the flux path, which yields more potential to reduce the active mass than RFPM generators. Figure 8 illustrates the active mass comparison of different 10 MW PM generators, where the RFPM generator serves as baseline and nine different TFPM generators are compared to it. The active mass of a claw pole design with limited pole area and up to four slots per phase seems to have potential against the baseline machine. To validate these analytical design results, three-dimensional finite element analyses are needed. 46 March 2011
Active mass [ton] 250 200 150 100 Generator total active mass Copper mass Stator core mass PM mass Rotor core mass Abbreviations U: U-core shape CP: claw poles 1..8: number of slots per phase L: limited axial length A: limited pole area 50 0 RFPM TFPM-U TFPM-CP/1/L TFPM-CP/2/L TFPM-CP/4/L TFPM-CP/8/L TFPM-CP/1/A TFPM-CP/2/A TFPM-CP/4/A TFPM-CP/8/A Generator type Figure 8: Active mass comparison of different 10 MW PM generators. Power electronics While acknowledging that doubly-fed induction generators are almost standard today, the UpWind research focused on full converter solutions for synchronous generators. Three different approaches to increase the power rating to the required level were analysed in detail. An example of a design approach including power device selection, selection of switching frequency, filter design, effi ciency curve, volume estimates and control scheme was drawn up. These concepts include the matrix converter, the three-level neutral point clamped (NPC) converter and the parallel interleaved converter. As a result of the benchmarks it can be noted that all topologies are potential candidates for next generation wind turbines and can serve the desired power conversion rating. The matrix converter is the topology offering the most potential for future developments. Currently the lack of tailor-made power semiconductors is a substantial drawback. The three-level NPC is a well established converter topology for the desired output power range, and it supports different output voltage levels. The parallel interleaved converter topology provides very good harmonic performance at the grid interface. The improvements on the harmonic performance are achieved at the expense of some additional circulating currents. A fact common to all three topologies is that the highest conversion effi ciency is achieved with the use of low voltage semiconductors (1700V IGBTs). However from the point of view of total system cost there is a tendency to aim for voltages that are as high as possible to decrease transformer and conductor costs. Design limits and solutions for very large wind turbines 47
Page 3 and 4: UpWind Design limits and solutions
Page 5 and 6: Contents 1. UpWind: Summary - a 20
Page 7 and 8: 4. UpWind: Research activities ....
Page 9 and 10: Contract number: 019945 (SES6) Dura
Page 11 and 12: WORK PACKAGES AUTHORS: Nicolas Fich
Page 13 and 14: “ They did not know it was imposs
Page 15 and 16: UpWind methodology - a lighthouse a
Page 17 and 18: Control and maintenance strategies
Page 19 and 20: 2008 250 m Ø 160 m Ø Rotor diamet
Page 21 and 22: In addition to defining a clear way
Page 23 and 24: UpWind: Scientific integration 1A1
Page 25 and 26: Results Subtask A: Reference wind t
Page 27 and 28: Subtask C: Development of (pre)stan
Page 29 and 30: “ The more your system is optimis
Page 31 and 32: In relation to the coming IEC stand
Page 33 and 34: References [1] Dahlberg J.A., et.al
Page 35 and 36: Results and conclusions First, prel
Page 37 and 38: 7. Electrical grid 7.1 Introduction
Page 39 and 40: UpWind: Technology integration 1B1:
Page 41 and 42: outboardblade Figure 5 : Sectional
Page 43 and 44: References [1] Ashwill T.D. and Paq
Page 45 and 46: [33] Schlotter M., “A Comparison
Page 47: unexpected failures is based on a b
Page 51 and 52: Research activities and working met
Page 53 and 54: Task 5: Interfaces Researching the
Page 55 and 56: • Capellaro M. and Kühn M., Aero
Page 57 and 58: This required a cost analysis of wi
Page 59 and 60: .UpWind: Research activities 2 Aero
Page 61 and 62: Engineering aerodynamic models by a
Page 63 and 64: “ Upscaling the current designs w
Page 65 and 66: Following the main route in impleme
Page 67 and 68: Figure 18: Damage modes according t
Page 69 and 70: “ UpWind demonstrated the importa
Page 71 and 72: Results Task 1: Integration of supp
Page 73 and 74: Task 3: Enhancement of design metho
Page 75 and 76: • Fischer, T., Kühn, M. and Pass
Page 77 and 78: Wind farms can be utilised to provi
Page 79 and 80: 18 Wind speed (upperline); rotor sp
Page 81 and 82: 25 v [m/s] 20 15 10 10 15 20 25 30
Page 83 and 84: 12 10 Blade 1 pitch angle [deg] 8 6
Page 85 and 86: References [1] Bossanyi E. Controll
Page 87 and 88: More specifically, the objectives o
Page 89 and 90: Figure 32: Commercial LIDARs underg
Page 91 and 92: “ UpWind showed the pathway for i
Page 93 and 94: 120 Break down Cyclic Condition bas
Page 95 and 96: References [1] Caselitz P., Giebhar
Page 97 and 98: Results An important task in UpWind
Page 99 and 100:
Typical wind farm layout optimisati
Page 101 and 102:
110% P/P N P [MW] adjustable power
Page 103 and 104:
The analysis of small island grids
Page 105:
Design limits and solutions for ver
show all

Design limits and solutions for very large wind turbines

Create successful ePaper yourself

Delete template?

Save as template?